Epidemiology of Sarcoidosis in Puerto Rico: A Population-Based Study From 2016 to 2018


 Background: Sarcoidosis is a systemic granulomatous and inflammatory disorder with significantvariability in ethnicity and geographical distribution. There is no descriptive data about theepidemiology of the disease among the geographically isolated Puerto Rican population.Objective: To estimate the incidence, prevalence, distribution, and healthcare burden ofsarcoidosis in a large nationwide population-based cohort in Puerto Rico.Methods: We conducted a descriptive and retrospective analysis using the health administrativeclaim database from the Puerto Rico Health Insurance Administration (ASES) from 2016 to 2018.The International Classification of Diseases-Tenth Revision coding (D86) was used for casedetection. Data on sex and age were used to estimate incidence and prevalence.Results: A total of 539 sarcoidosis cases were identified over the 3-year study. The median ageat diagnosis was 59 years old. The average annual incidence rate was 9.4/100,000. The baselineprevalence was 15.4 per 100,000. Females represented the 67.5% of the cases. The frequency ofsarcoidosis in women was higher than in males starting at the age range 18-34 years (GLMp<0.03). Patients with lung and unspecified sarcoidosis predominates with 37% and 32% of cases,respectively.Conclusions: Annual incidence rate of sarcoidosis in Island Puerto Ricans is among the highestin a single geographically isolated ethnic group reported globally. The overall mean age ofindividuals with sarcoidosis in Puerto Rico represent the oldest among previous epidemiologicalsurveys conducted worldwide.


Introduction
Sarcoidosis is a chronic multisystemic disorder of unknown etiology characterized by the presence of granulomatous inflammatory lesions most commonly involving the lungs and adjacent lymph nodes (1). The liver, skin, eyes, spleen, salivary glands, heart, nervous system, muscles, bones, and other organs may also be involved (2). The disease can range from an acute self-limiting process to progressive chronic scarring which may permanently impair organ function.
Once considered a rare disease, sarcoidosis is now known to occur globally, affecting both sexes and all ethnicities and ages (3). Worldwide geographical and demographic variations have been observed in sarcoidosis occurrence. The highest incidence and prevalence rates are found in Northern Europeans and African Americans, with an average onset between 47 and 51 years old (3). Literature review consistently describes Puerto Rican Hispanics living on the U.S mainland as one of the ethnic groups to be at a higher risk of developing the disease (4,5,6,23). Such reports have indicated that sarcoidosis in Puerto Rican Hispanics involves only lungs and skin (5,6).
However, there are no epidemiological studies to clarify better the heterogeneity of the disease on the island of Puerto Rico.
The availability of nationwide healthcare databases allows for specific information regarding the incidence and prevalence of a disease across an entire population for individuals with healthcare insurance for a specific period (7,8). In addition, it allows researchers to compare disparities across racial / ethnic groups. The databases' access to identify sarcoidosis may provide the healthcare systems with tools to assess the human and economic costs of the disease, improve efficiency, and optimize management. Also, it has an advantage that members may be more likely to follow through with the tests and procedures needed to confirm the diagnosis.
Approximately half of Puerto Rico's 3.2 million residents have low incomes and depend upon the public health system for their medical care (9). The Puerto Rico Department of Health made a cooperative agreement with the Puerto Rico Health Insurance Administration (PRHIA) also known as Administración de Seguros de Salud de Puerto Rico (ASES), to administer the island-wide universal health insurance system (10). Since 1993, ASES has the responsibility of implementing, managing, and negotiating through contracts with insurers and health service organizations to offer services to approximately 1.5 million indigent medical Puerto Ricans. There are two main healthcare systems administered by ASES: Mi Salud (VITAL or PSG-Plan de Salud del Gobierno), which provides healthcare services to over a million people across the island and the Medicare "Platino" with approximately 250,000 patients (10).
The purpose of this study was to estimate the incidence, prevalence, distribution and healthcare burden of sarcoidosis in a large nationwide population-based cohort in Puerto Rico, using the available access to claims data from ASES.

Data Source and Collection
An analytical file of patients with sarcoidosis was done using the ASES encrypted database, which represents approximately 1.6 million patients annually. The database contains information regarding patient age, sex, enrollment history, medical diagnoses, procedures performed and payment amounts for the time period spanning January 2016 to December 2018 We selected patients living in Puerto Rico who filed medical claims from two government healthcare plans ("PSG" and "Platino") under the D86 code "sarcoidosis" of the International

Statistical analyses:
Categorical variables were compared using simulated Chi-square test. To analyze the tendency of the prevalence of sex by age, the generalized lineal model analysis was performed.
The risk estimates of developing sarcoidosis in women with an increase of age was calculated from a Poisson regression model using the generalized linear model (12). All analysis and graphics were produced using ggplot R version 3.6.0 (13). The frequency of new cases was calculated with comparedDF package program (14). We performed an ordered Generalized Linear Model (12) to evaluate if the change in frequency by sex and age were similar.

Ethics
This is a descriptive, non-interventional study involving the use of unidentified data.

Nevertheless, it was revised and approved by the Institutional Review Board of the San Juan
Bautista School of Medicine (IRB # EMSJBIRB-16-2019).

Baseline characteristics of patients with sarcoidosis in Puerto Rico 2016-2018
Over the 3-year study period, we identified 539 unique individuals which have been  Table 1).
The age and sex distribution of patients with sarcoidosis in Puerto Rico from 2016 to 2018 is shown in Figure 1a. Individuals from 0-17 years of age showed the lowest frequency of patients with sarcoidosis. Women 65+ years represent the highest frequency of patients with sarcoidosis.
The frequency of sarcoidosis in women was higher than in males starting at the age range 18-34 years ( Figure 1a). While the male frequency rate reached a plateau in the 55-64-yr-old group, females showed a peak at 65+ year-old ( Figure 1a). To test if the differences in the frequency of sarcoidosis by age range and sex are significant, we used an Ordered Generalized Linear Model.
We used a model including the age group category ("0-17", "18-34","35-44" ,"45-54" ,"55-64" ,"65+"), sex and the interaction between both the variables. The result show that both age range and sex influence the frequency of sarcoidosis, being the rate of change in frequency in sarcoidosis much higher in females than in males (Ordered GLM p < 0.03) (Figure 1b).

Organ involvement in sarcoidosis
We evaluated the frequency of specific organ involvement of sarcoidosis in Puerto Rico We next compared the characteristics of the patients by sex and organ involvement ( Figure   2b). Lung, unspecified, other sites, and skin sarcoidosis were more frequent in females than in males but not statistically significant (Pearson's Chi squared test= 5.96; df =3; p = 0.11).

Comorbidities
We evaluated the frequency and types of the comorbidities that accompany sarcoidosis in Puerto Rico under the period of study. Regression analysis results showed a clear correlation between an increase in the number of comorbidities with age ( Figure 3a). Statistically significant differences in the number of comorbidities between women and men is observed for ages before 50 ( Figure 3a). Figure 3b presents the diseases most frequently reported as comorbidity in sarcoidosis patients. Of the 493 patients with associated comorbidities, the five most common diagnosis was hypertension in 69% of patients, followed by Genito urinary disorders (67.7%), Diabetes type 2 (53.5%), Hyperlipidemia (48.5%) and Anemia (45%). In addition to each patient being assigned an ICD-10 code associated with Sarcoidosis, patients had multiple other comorbidities (mean of 6.2).

Healthcare resources and cost of sarcoidosis care in Puerto Rico
We assessed the healthcare resources and cost of sarcoidosis care to government insurance payers in Puerto Rico during the 2016 to 2018 period of study ( year-old), which is similar to that reported from North European populations (19).
Sarcoidosis is consistently reported to show a monophasic pattern of incidence, with peak between 42 and 55 years of age (3). The peak age in the USA group was 35-45 yrs (22), whereas in Japan the peak showed a biphasic pattern (20-34-yr-old in males and 50-64-yr-old in females;17). In North Europeans, the age at diagnosis in men was 45 compared with 54 in women (19). In Korea, the mean age of patients with sarcoidosis was 48.9 yrs (18). The peak ages of sarcoidosis in our study (mean = 59 years old; male = 55-64-year-old, female = +65-year-old) represent the oldest reported. This finding is strongly influenced by the high proportion of elderly patients living in Puerto Rico (24). The results suggest that sarcoidosis in Puerto Rico is concentrated at a middle to advanced age. The economic impact of sarcoidosis within this group may be more significant than any other age group due to factors such as: loss of working capacity and income, productivity and exacerbations with co-morbidities associated and not associated with the condition itself.
Pulmonary involvement is highly prevalent in sarcoidosis, as it was reported in several studies worldwide (22,23,26,27). However, organ involvement in sarcoidosis have been shown to vary among different geographic and ethnic groups. For example, ocular and cardiac manifestation of sarcoidosis is more frequent among Japanese (17). Skin sarcoidosis is prevalent among African Americans but uncommon among Caucasian Population (22,25). In our study, the predominant organ involved was the lung with a prevalence of 37%. However, our prevalence was lower than those reported in the USA (95%; 22), Mexico (65%; 26) and Poland (79.2%; 27) studies.
Prior studies had reported that Puerto Rican patients have a high frequency of skin sarcoidosis (5,19,23). However, we did not find that high frequency of skin sarcoidosis among Island Puerto Ricans since our estimate is relatively low (6.6%). According to Soto-Gomez et al. (28), worldwide prevalence of skin sarcoidosis is 25%.
In our study, unspecified sarcoidosis was the second most common sarcoidosis found with a prevalence of thirty-two (32%) of patients. This contrast with a retrospective study made in Poland in which only 5% of patients were diagnosed with unspecified sarcoidosis (27). We can only speculate about the cause of the frequent use of "unspecified sarcoidosis" code in Puerto Rico.
One of the possible reasons may be attributed to differences in the level, and expertise of both knowledge and diagnostic techniques on both studies. In the Poland cohort study, all patients were The goals of therapy in sarcoidosis are to reduce the inflammation of the affected tissues, reduce the impact of granuloma development, and prevent the development of lung fibrosis and another irreversible organ damage (30). The use of corticosteroids is considered the first line of therapy for the treatment of inflammation associated with sarcoidosis (30,31,32). The most prescribed corticosteroid in our study was prednisone (42%). This is slightly lower than the estimates reported by Baughman et al (7) who found that 49.1% of new and 56% of continuing cases of sarcoidosis patients enrolled in a national health plan in US were being treated with the drug. The antimalarial drug hydroxychloroquine (Plaquenil) was the second most used drug in our study (20%), as it was in the study of Baughman et al (7). This agent has been reported to be effective in people who have sarcoidosis of skin, lung and nervous system or a high level of calcium in their blood. However, our findings contrast with findings from Bradford-Rice (20% vs 4.1% on lower cost patients and 20% vs 6.8% on high cost patients, 33). It is known that the prolonged use of prednisone and hydroxychloroquine is associated with the development of glaucoma and cataracts (30) and this may explain the 8.1% in expenses associated with the use of prescription drugs for ophthalmic use found in our study (  (Figure 3b). The prevalence of hypertension and Diabetes mellitus in our study is similar to that reported from population-based studies in Puerto Rico (34,35). In addition, the frequent use of prednisone in our study may contribute for the high prevalence of these two comorbidities since hypertension and diabetes mellitus is known to be a side-effect of prednisone (30) The heavy load to the health care expenses and patient well-being associated with the somatic manifestations and psychosocial, economic and comorbid conditions of sarcoidosis are well known (7,8 insurance and those without insurance coverage. Second, data on the incidence and prevalence of sarcoidosis in Puerto Rico might have been overestimated over the study period due to possible inaccuracies in the use of ICD codes assigned by the health care providers to diagnose sarcoidosis and subtypes. Since sarcoidosis may mimic other chronic inflammatory / granulomatous conditions, the risk of misclassification or over diagnosis of sarcoidosis cannot be excluded. We were unable to confirm the sarcoidosis cases through the access of medical records data (i.e., imaging, pathologic results). Third, there is a possibility that we may oversample an older group of subjects in determining the incidence and prevalence of sarcoidosis in Puerto Rico since the population in the ASES database is older than the general population in Puerto Rico.
Despite the limitations, the large sample size obtained by the use of the insurance claims database from the Puerto Rico health insurance and Medicare can be considered as an advantage, since it provided data on a large subset of the population and a first estimates of prevalence and incidence of sarcoidosis in Puerto Rico.

Conclusions
The data obtained in the present study shows differences in age, sex, and organ